Skip to contentSkip to conversation

extensions: libxt_multiport: Add translation to nft

PATCH Superseded Laura Garcia Liebana 3 participants 7 messages May 30, 2016 – May 31, 2016 netfilter

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 extensions: libxt_multiport: Add translation to nft

Add translation for multiport to nftables, which it's supported natively.

Examples:

$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80,81 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80,81} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80-88} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport ! --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport != 80-88 counter accept

---
Changes in v2:
- Add curley brackets to lists and range of ports.

 extensions/libxt_multiport.c | 116 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 116 insertions(+)
extensions/libxt_multiport.c +116 -0
index 03af5a9..25b5589 100644
@@ -468,6 +468,118 @@ static void multiport_save6_v1(const void *ip_void,
 	__multiport_save_v1(match, ip->proto);
 }
 
+static int __multiport_xlate(const void *ip, const struct xt_entry_match *match,
+			     struct xt_xlate *xl, int numeric)
+{
+	const struct xt_multiport *multiinfo
+		= (const struct xt_multiport *)match->data;
+	unsigned int i;
+
+	switch (multiinfo->flags) {
+	case XT_MULTIPORT_SOURCE:
+		xt_xlate_add(xl, "sport ");
+		break;
+	case XT_MULTIPORT_DESTINATION:
+		xt_xlate_add(xl, "dport ");
+		break;
+	case XT_MULTIPORT_EITHER:
+		return 0;
+	}
+
+	if (multiinfo->count > 1)
+		xt_xlate_add(xl, "{ ");
+
+	for (i = 0; i < multiinfo->count; i++)
+		xt_xlate_add(xl, "%s%u", i ? "," : "", multiinfo->ports[i]);
+
+	if (multiinfo->count > 1)
+		xt_xlate_add(xl, "}");
+
+	xt_xlate_add(xl, " ");
+
+	return 1;
+}
+
+static int multiport_xlate(const void *ip, const struct xt_entry_match *match,
+			   struct xt_xlate *xl, int numeric)
+{
+	uint8_t proto = ((const struct ipt_ip *)ip)->proto;
+
+	xt_xlate_add(xl, "%s ", proto_to_name(proto));
+	return __multiport_xlate(ip, match, xl, numeric);
+}
+
+static int multiport_xlate6(const void *ip, const struct xt_entry_match *match,
+			    struct xt_xlate *xl, int numeric)
+{
+	uint8_t proto = ((const struct ip6t_ip6 *)ip)->proto;
+
+	xt_xlate_add(xl, "%s ", proto_to_name(proto));
+	return __multiport_xlate(ip, match, xl, numeric);
+}
+
+static int __multiport_xlate_v1(const void *ip,
+				const struct xt_entry_match *match,
+				struct xt_xlate *xl, int numeric)
+{
+	const struct xt_multiport_v1 *multiinfo
+		= (const struct xt_multiport_v1 *)match->data;
+	unsigned int i;
+
+	switch (multiinfo->flags) {
+	case XT_MULTIPORT_SOURCE:
+		xt_xlate_add(xl, "sport ");
+		break;
+	case XT_MULTIPORT_DESTINATION:
+		xt_xlate_add(xl, "dport ");
+		break;
+	case XT_MULTIPORT_EITHER:
+		return 0;
+	}
+
+	if (multiinfo->invert) {
+		xt_xlate_add(xl, "!= ");
+	} else {
+		if (multiinfo->count > 1)
+			xt_xlate_add(xl, "{ ");
+	}
+
+	for (i = 0; i < multiinfo->count; i++) {
+		xt_xlate_add(xl, "%s%u", i ? "," : "", multiinfo->ports[i]);
+		if (i && multiinfo->invert)
+			return 0;
+		if (multiinfo->pflags[i])
+			xt_xlate_add(xl, "-%u", multiinfo->ports[++i]);
+	}
+
+	if (multiinfo->count > 1 && !multiinfo->invert)
+		xt_xlate_add(xl, "}");
+
+	xt_xlate_add(xl, " ");
+
+	return 1;
+}
+
+static int multiport_xlate_v1(const void *ip,
+			      const struct xt_entry_match *match,
+			      struct xt_xlate *xl, int numeric)
+{
+	uint8_t proto = ((const struct ipt_ip *)ip)->proto;
+
+	xt_xlate_add(xl, "%s ", proto_to_name(proto));
+	return __multiport_xlate_v1(ip, match, xl, numeric);
+}
+
+static int multiport_xlate6_v1(const void *ip,
+			       const struct xt_entry_match *match,
+			       struct xt_xlate *xl, int numeric)
+{
+	uint8_t proto = ((const struct ip6t_ip6 *)ip)->proto;
+
+	xt_xlate_add(xl, "%s ", proto_to_name(proto));
+	return __multiport_xlate_v1(ip, match, xl, numeric);
+}
+
 static struct xtables_match multiport_mt_reg[] = {
 	{
 		.family        = NFPROTO_IPV4,
@@ -482,6 +594,7 @@ static struct xtables_match multiport_mt_reg[] = {
 		.print         = multiport_print,
 		.save          = multiport_save,
 		.x6_options    = multiport_opts,
+		.xlate         = multiport_xlate,
 	},
 	{
 		.family        = NFPROTO_IPV6,
@@ -496,6 +609,7 @@ static struct xtables_match multiport_mt_reg[] = {
 		.print         = multiport_print6,
 		.save          = multiport_save6,
 		.x6_options    = multiport_opts,
+		.xlate         = multiport_xlate6,
 	},
 	{
 		.family        = NFPROTO_IPV4,
@@ -510,6 +624,7 @@ static struct xtables_match multiport_mt_reg[] = {
 		.print         = multiport_print_v1,
 		.save          = multiport_save_v1,
 		.x6_options    = multiport_opts,
+		.xlate         = multiport_xlate_v1,
 	},
 	{
 		.family        = NFPROTO_IPV6,
@@ -524,6 +639,7 @@ static struct xtables_match multiport_mt_reg[] = {
 		.print         = multiport_print6_v1,
 		.save          = multiport_save6_v1,
 		.x6_options    = multiport_opts,
+		.xlate         = multiport_xlate6_v1,
 	},
 };
 
Signature · 1 line
2.7.0
#1 lore raw
Quoted 4 lines · Laura Garcia Liebana

Add translation for multiport to nftables, which it's supported natively.

Examples:

Quoted 9 lines
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80,81 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80,81} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80-88} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport ! --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport != 80-88 counter accept

Lets clarify the syntax, this is valid:

tcp dport 8000-8100
tcp dport { 8000-8100}
tcp dport { 8000-8100, 9000-9100}

but they mean different things. It seems we should avoid the braces {} for the range case, otherwise we would be using a set with a single element.

However,

tcp dport {8000,8100} <-- valid
tcp dport 8000,8100 <-- invalid

So we should always use braces {} in the non-range case.

Same seems to apply in the case of inversion.

so we end with this combinations:

tcp dport {x,x}
tcp dport != {x,x}
tcp dport x-x
tcp dport != x-x

BTW, related to this, there seems to be a bug in nftables:

% nft add rule t c tcp dport != {80, 81}

BUG: invalid expression type set

nft: evaluate.c:1463: expr_evaluate_relational: Assertion `0' failed. Aborted

Quoted 23 lines

---
Changes in v2:
- Add curley brackets to lists and range of ports.

 extensions/libxt_multiport.c | 116 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 116 insertions(+)
extensions/libxt_multiport.c +7 -0
index 03af5a9..25b5589 100644
@@ -468,6 +468,118 @@ static void multiport_save6_v1(const void *ip_void,
        __multiport_save_v1(match, ip->proto);
 }
 
+static int __multiport_xlate(const void *ip, const struct xt_entry_match *match,
+                            struct xt_xlate *xl, int numeric)
+{
+       const struct xt_multiport *multiinfo
+               = (const struct xt_multiport *)match->data;
+       unsigned int i;
+

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 487

Quoted 10 lines
extensions/libxt_multiport.c +10 -0
+       switch (multiinfo->flags) {
+       case XT_MULTIPORT_SOURCE:
+               xt_xlate_add(xl, "sport ");
+               break;
+       case XT_MULTIPORT_DESTINATION:
+               xt_xlate_add(xl, "dport ");
+               break;
+       case XT_MULTIPORT_EITHER:
+               return 0;
+       }

this case XT_MULTIPORT_EITHER seems unsupported in nftables.

Is there anything established to do in case we find an impossible translation? print a warning or something? I don't know right now. I guess we should avoid printing an invalid nftables rule as if the translation was 100% ok (which is not true in this case).

Just wondering, I should check myself because I don't know this right now.

Quoted 41 lines
extensions/libxt_multiport.c +41 -0
+
+       if (multiinfo->count > 1)
+               xt_xlate_add(xl, "{ ");
+
+       for (i = 0; i < multiinfo->count; i++)
+               xt_xlate_add(xl, "%s%u", i ? "," : "", multiinfo->ports[i]);
+
+       if (multiinfo->count > 1)
+               xt_xlate_add(xl, "}");
+
+       xt_xlate_add(xl, " ");
+
+       return 1;
+}
+
+static int multiport_xlate(const void *ip, const struct xt_entry_match *match,
+                          struct xt_xlate *xl, int numeric)
+{
+       uint8_t proto = ((const struct ipt_ip *)ip)->proto;
+
+       xt_xlate_add(xl, "%s ", proto_to_name(proto));
+       return __multiport_xlate(ip, match, xl, numeric);
+}
+
+static int multiport_xlate6(const void *ip, const struct xt_entry_match *match,
+                           struct xt_xlate *xl, int numeric)
+{
+       uint8_t proto = ((const struct ip6t_ip6 *)ip)->proto;
+
+       xt_xlate_add(xl, "%s ", proto_to_name(proto));
+       return __multiport_xlate(ip, match, xl, numeric);
+}
+
+static int __multiport_xlate_v1(const void *ip,
+                               const struct xt_entry_match *match,
+                               struct xt_xlate *xl, int numeric)
+{
+       const struct xt_multiport_v1 *multiinfo
+               = (const struct xt_multiport_v1 *)match->data;
+       unsigned int i;
+

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 538

Quoted 10 lines
extensions/libxt_multiport.c +10 -0
+       switch (multiinfo->flags) {
+       case XT_MULTIPORT_SOURCE:
+               xt_xlate_add(xl, "sport ");
+               break;
+       case XT_MULTIPORT_DESTINATION:
+               xt_xlate_add(xl, "dport ");
+               break;
+       case XT_MULTIPORT_EITHER:
+               return 0;
+       }

same XT_MULTIPORT_EITHER here.

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 545

Quoted 7 lines
extensions/libxt_multiport.c +7 -0
+
+       if (multiinfo->invert) {
+               xt_xlate_add(xl, "!= ");
+       } else {
+               if (multiinfo->count > 1)
+                       xt_xlate_add(xl, "{ ");
+       }

This if/else seems bogus. We only allow port sets if not inverting?

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 550

Quoted 5 lines
extensions/libxt_multiport.c +5 -0
+
+       for (i = 0; i < multiinfo->count; i++) {
+               xt_xlate_add(xl, "%s%u", i ? "," : "", multiinfo->ports[i]);
+               if (i && multiinfo->invert)
+                       return 0;

This return here could mean that we build an incomplete nftables rule (ie, missing '}')

Quoted 11 lines
extensions/libxt_multiport.c +11 -0
+               if (multiinfo->pflags[i])
+                       xt_xlate_add(xl, "-%u", multiinfo->ports[++i]);
+       }
+
+       if (multiinfo->count > 1 && !multiinfo->invert)
+               xt_xlate_add(xl, "}");
+
+       xt_xlate_add(xl, " ");
+
+       return 1;
+}
Signature · 5 lines
Arturo Borrero González
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw
Quoted 41 lines · Arturo Borrero Gonzalez
Quoted 4 lines · Laura Garcia Liebana

Add translation for multiport to nftables, which it's supported natively.

Examples:

Quoted 9 lines
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80,81 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80,81} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80-88} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport ! --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport != 80-88 counter accept

Lets clarify the syntax, this is valid:

tcp dport 8000-8100
tcp dport { 8000-8100}
tcp dport { 8000-8100, 9000-9100}

but they mean different things. It seems we should avoid the braces {} for the range case, otherwise we would be using a set with a single element.

However,

tcp dport {8000,8100} <-- valid
tcp dport 8000,8100 <-- invalid

So we should always use braces {} in the non-range case.

Same seems to apply in the case of inversion.

so we end with this combinations:

tcp dport {x,x}
tcp dport != {x,x}
tcp dport x-x
tcp dport != x-x

Quoted 6 lines

BTW, related to this, there seems to be a bug in nftables:

% nft add rule t c tcp dport != {80, 81}

BUG: invalid expression type set

nft: evaluate.c:1463: expr_evaluate_relational: Assertion `0' failed. Aborted

This is not yet supported. This requires a small kernel patch to allow inversions in the nft_lookup.c. Then, the little extra code for libnftnl and nft.

All tests for this usecase are disabled at the moment.

#5 lore raw
In reply to Arturo Borrero Gonzalez
Quoted 17 lines · Arturo Borrero Gonzalez
Quoted 4 lines · Laura Garcia Liebana

Add translation for multiport to nftables, which it's supported natively.

Examples:

Quoted 9 lines
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80,81 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80,81} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport { 80-88} counter accept
$ sudo iptables-translate -t filter -A INPUT -p tcp -m multiport ! --dports 80:88 -j ACCEPT
nft add rule ip filter INPUT ip protocol tcp tcp dport != 80-88 counter accept

Lets clarify the syntax, this is valid:

Quoted 8 lines

tcp dport 8000-8100
tcp dport { 8000-8100}
tcp dport { 8000-8100, 9000-9100}

but they mean different things. It seems we should avoid the braces {} for the range case, otherwise we would be using a set with a single element.

Yes, you're right. I'll change it in order to allow port ranges without {}.

Quoted 5 lines

However,

tcp dport {8000,8100} <-- valid
tcp dport 8000,8100 <-- invalid

Quoted 8 lines

So we should always use braces {} in the non-range case.

Same seems to apply in the case of inversion.

so we end with this combinations:

tcp dport {x,x}
tcp dport != {x,x}

This is not supported.

Quoted 47 lines

tcp dport x-x
tcp dport != x-x

BTW, related to this, there seems to be a bug in nftables:

% nft add rule t c tcp dport != {80, 81}

BUG: invalid expression type set

nft: evaluate.c:1463: expr_evaluate_relational: Assertion `0' failed. Aborted

Quoted 23 lines

---
Changes in v2:
- Add curley brackets to lists and range of ports.

 extensions/libxt_multiport.c | 116 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 116 insertions(+)
extensions/libxt_multiport.c +7 -0
index 03af5a9..25b5589 100644
@@ -468,6 +468,118 @@ static void multiport_save6_v1(const void *ip_void,
        __multiport_save_v1(match, ip->proto);
 }
 
+static int __multiport_xlate(const void *ip, const struct xt_entry_match *match,
+                            struct xt_xlate *xl, int numeric)
+{
+       const struct xt_multiport *multiinfo
+               = (const struct xt_multiport *)match->data;
+       unsigned int i;
+

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 487

Quoted 10 lines
extensions/libxt_multiport.c +10 -0
+       switch (multiinfo->flags) {
+       case XT_MULTIPORT_SOURCE:
+               xt_xlate_add(xl, "sport ");
+               break;
+       case XT_MULTIPORT_DESTINATION:
+               xt_xlate_add(xl, "dport ");
+               break;
+       case XT_MULTIPORT_EITHER:
+               return 0;
+       }

this case XT_MULTIPORT_EITHER seems unsupported in nftables.

Quoted 5 lines

Is there anything established to do in case we find an impossible translation? print a warning or something? I don't know right now. I guess we should avoid printing an invalid nftables rule as if the translation was 100% ok (which is not true in this case).

It was agreed to return 0 if the translation is not supported for nftables. Currently, it does.

Quoted 48 lines

Just wondering, I should check myself because I don't know this right now.

Quoted 45 lines
extensions/libxt_multiport.c +45 -0
+
+       if (multiinfo->count > 1)
+               xt_xlate_add(xl, "{ ");
+
+       for (i = 0; i < multiinfo->count; i++)
+               xt_xlate_add(xl, "%s%u", i ? "," : "", multiinfo->ports[i]);
+
+       if (multiinfo->count > 1)
+               xt_xlate_add(xl, "}");
+
+       xt_xlate_add(xl, " ");
+
+       return 1;
+}
+
+static int multiport_xlate(const void *ip, const struct xt_entry_match *match,
+                          struct xt_xlate *xl, int numeric)
+{
+       uint8_t proto = ((const struct ipt_ip *)ip)->proto;
+
+       xt_xlate_add(xl, "%s ", proto_to_name(proto));
+       return __multiport_xlate(ip, match, xl, numeric);
+}
+
+static int multiport_xlate6(const void *ip, const struct xt_entry_match *match,
+                           struct xt_xlate *xl, int numeric)
+{
+       uint8_t proto = ((const struct ip6t_ip6 *)ip)->proto;
+
+       xt_xlate_add(xl, "%s ", proto_to_name(proto));
+       return __multiport_xlate(ip, match, xl, numeric);
+}
+
+static int __multiport_xlate_v1(const void *ip,
+                               const struct xt_entry_match *match,
+                               struct xt_xlate *xl, int numeric)
+{
+       const struct xt_multiport_v1 *multiinfo
+               = (const struct xt_multiport_v1 *)match->data;
+       unsigned int i;
+
+       switch (multiinfo->flags) {
+       case XT_MULTIPORT_SOURCE:
+               xt_xlate_add(xl, "sport ");
+               break;
Quoted 9 lines

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 487

Quoted 6 lines
extensions/libxt_multiport.c +6 -0
+       case XT_MULTIPORT_DESTINATION:
+               xt_xlate_add(xl, "dport ");
+               break;
+       case XT_MULTIPORT_EITHER:
+               return 0;
+       }

same XT_MULTIPORT_EITHER here.

Same as before.

Quoted 10 lines

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 545

Quoted 7 lines
extensions/libxt_multiport.c +7 -0
+
+       if (multiinfo->invert) {
+               xt_xlate_add(xl, "!= ");
+       } else {
+               if (multiinfo->count > 1)
+                       xt_xlate_add(xl, "{ ");
+       }

This if/else seems bogus. We only allow port sets if not inverting?

This should be now changed as we've to accept port ranges without {}.

Quoted 9 lines

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 550

Quoted 5 lines
extensions/libxt_multiport.c +5 -0
+
+       for (i = 0; i < multiinfo->count; i++) {
+               xt_xlate_add(xl, "%s%u", i ? "," : "", multiinfo->ports[i]);
+               if (i && multiinfo->invert)
+                       return 0;

This return here could mean that we build an incomplete nftables rule (ie, missing '}')

Such condition cares that there is no translation in nftables for:

tcp dport != { 80,88 }

So the closed } doesn't even matter.

Quoted 4 lines

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 554

Quoted 4 lines
extensions/libxt_multiport.c +4 -0
+               if (multiinfo->pflags[i])
+                       xt_xlate_add(xl, "-%u", multiinfo->ports[++i]);
+       }
+

If the rule needs such closed } , then it's controlled here:

Quoted 11 lines
Quoted 7 lines
extensions/libxt_multiport.c +7 -0
+       if (multiinfo->count > 1 && !multiinfo->invert)
+               xt_xlate_add(xl, "}");
+
+       xt_xlate_add(xl, " ");
+
+       return 1;
+}
Signature · 1 line
Arturo Borrero González

--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#6 lore raw
Quoted 10 lines · Laura Garcia
Quoted 9 lines

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 550

Quoted 5 lines
extensions/libxt_multiport.c +5 -0
+
+       for (i = 0; i < multiinfo->count; i++) {
+               xt_xlate_add(xl, "%s%u", i ? "," : "", multiinfo->ports[i]);
+               if (i && multiinfo->invert)
+                       return 0;

This return here could mean that we build an incomplete nftables rule (ie, missing '}')

Quoted 6 lines

Such condition cares that there is no translation in nftables for:

tcp dport != { 80,88 }

So the closed } doesn't even matter.

Then, at the start of the xlate function, I would suggest to do an early return if we are about to translate an unsupported rule.

No need to keep checking if we are in the inversion case

Quoted 3 lines
Quoted 3 lines

extensions/libxt_multiport.c static void multiport_save6_v1(const void *ip_void, · line 557

Quoted 3 lines
extensions/libxt_multiport.c +3 -0
+       if (multiinfo->count > 1 && !multiinfo->invert)
+               xt_xlate_add(xl, "}");
+

like here

Signature · 5 lines
Arturo Borrero González
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#3 lore raw
In reply to Pablo Neira Ayuso
Quoted 5 lines · Pablo Neira Ayuso

This is not yet supported. This requires a small kernel patch to allow inversions in the nft_lookup.c. Then, the little extra code for libnftnl and nft.

All tests for this usecase are disabled at the moment.

What is your idea of the implementation?

Perhaps adding NFTA_LOOKUP_FLAGS and also:
enum nft_lookup_flags {
NFT_LOOKUP_F_INV = (1 << 0),
};

Signature · 5 lines
Arturo Borrero González
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#4 lore raw
Quoted 5 lines · Arturo Borrero Gonzalez
Quoted 4 lines · Pablo Neira Ayuso

This is not yet supported. This requires a small kernel patch to allow inversions in the nft_lookup.c. Then, the little extra code for libnftnl and nft.

Quoted 8 lines
Quoted 1 line

All tests for this usecase are disabled at the moment.

What is your idea of the implementation?

Perhaps adding NFTA_LOOKUP_FLAGS and also:

enum nft_lookup_flags {
   NFT_LOOKUP_F_INV = (1 << 0),
};

Something like that, yes.